Dear amber users,
I have just technical question related to QM->ff fitting using Amber.
Let say that I want to derive some ff parameters from QM
using just comparison of QM and ff relative energies of the set
of small distorted molecular segments of my interest.
This approach could be sumarised this way:
#1
Prepare set of several distorted configurations of the given molecule
(X1,X2,X3, ....Xn).
#2
Calculate QM minimal structure (structure after QM energy minimisation )
Xmin_QM and it's QM energy E_QM(Xmin_QM)
and also the energies E_QM(X1), E_QM(X2), E_QM(X3) ...
#3
Calculate QM relative energies Eri_QM = E_QM(Xi)- E_QM(Xmin_QM) ( +
convert units )
#4
Calculate FF minimal structure (structure after FF energy minimisation )
Xmin_FF and it's FF energy E_FF(Xmin_FF)
and also the FF energies E_FF(X1), E_FF(X2), E_FF(X3) ...
#5
Calculate FF relative energies Eri_FF = E_FF(Xi)- E_FF(Xmin_FF)
#6
Calculate differences between QM and FF relative energies ( dEi = Eri_QM -
Eri_FF )
#7
If the sum of dEi^2 is greater than desired value, adjust properly the FF
parameters of the interest and
repeat the procedure from the step #4.
OK, this is clear but to be able to do this whole procedure effectively
it is necessary implement
all the steps #4-#7 in one program or script and not to do it
interactively by hand (i.e. load by hand X1,X2 ...Xn
into tleap from source PDBs together with PREPIN and relevant FRCMOD file
containing the actual FF parameters of the interest and creat PRMTOP and
INPCRD files,
start manually FF (Sander) minimisation and FF energy calculation ...)
I think this could be with some small effort done using just some parts of
the tleap source code and mix it with my own code.
So the most crucial thing is to be able identify the relevant source parts
of the tleap package.
Here it means at least parts responsible for loading PDBs structures and
FRCMOD and PREPIN files and
also parts for creating of PRMTOP and INCPRD files. This is important
since step #7 could in the simplest
way involve repetitive modifications of FRCMOD file and so consequently
repetitive loading of all structures into tleap and
repetitive creation of the PRMTOP and INPCRD input files. Of course if one
understand how the FF data are managed in
the tleap source code (understanding of used structures, arrays) one can
avoid repetitive FRCMOD file modifications and
structures loading, but it really needs detail information about FF
parameters representation inside the tleap code.
But anyway still it is here the need for the repetitive PRMTOP and INPCRD
files creation, of course it is detail to
implement in own program repetitive loading of Sander calculation and
proper processing of the output data.
So the crucial points here are in my opinion at least to identify all the
tleap source code parts/functions for loading structures and PREPIN,
FRCMOD files
and for PRMTOP and INPCRD files creation.
After first look into leap source dir. it is clear to me that the part
responsible for the reading of the PDB structures is
written in pdb_read.c . I would really appreciate help in identification
of the other parts (files, functions) responsible for above
mentioned actions as well as for all the relevant comments to this topic.
Thanks in advance for any relevant information !
Best,
Marek
--
Tato zpráva byla vytvořena převratným poštovním klientem Opery:
http://www.opera.com/mail/
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Nov 05 2009 - 14:00:02 PST
